Alright well I think it's working as well as it ever will. I don't necessarily need to go into all the details on the theory of this project, as Ben Heck covers it pretty well in his video. But I am happy to share what I learned:
1) the .030" default drill size used for the default pin headers footprints in KiCAD is just a bit to small (at least on boards made by OSHPark). All my other components were fine but my pin headers could have been .035, maybe bigger even.
2) You need to double check the keep out zones on your footprints per IC. One of my IC's was only as long as the pins, but another stuck out about 15 or 20 mills. It's a miracle that my pots still fit.
3) I think I know why my setup is having weird issues with flashing and off colors. On the WS2812 datasheet, it specified timing tolerances for high time and low time for both the 0 and 1 pulse. So it listed four times, each with a tolerance of ±150 nanoseconds. I took this to mean that it could run at a frequency of between 625 kHz and 1.17 MHz, but the datasheet also specifies on 800 kHz data transmission frequency. Oddly enough, it specifies the nominal 0 code as taking 1150 ns but the nominal 1 code taking 1300 ns. But I think that because I am not transmitting very close to 800kHz, the end of my transmission is being skewed slightly. I will probably try and design a V2 of the board, with its own onboard clock, and see if that fixes these issues.